The authors generated a suite of engineered Htr2a mouse lines—including an Htr2a‑EGFP‑CT‑IRES‑CreERT2 reporter, a humanised Htr2a line and a constitutive Htr2a‑Cre line—that enable independent visualisation of HTR2A receptors and HTR2A‑expressing cells and provide detailed anatomical and cell‑type maps. They validated the lines with psychedelic‑induced behavioural phenotypes and electrophysiology showing extracellular 5‑HT produces an HTR2A‑mediated increase in firing of genetically identified pyramidal neurons, consistent with plasma‑membrane localisation and utility for in vivo studies of psychedelic drug actions.
Psychedelic drugs like lysergic acid diethylamide (LSD) and psilocybin have emerged as potentially transformative therapeutics for many neuropsychiatric diseases, including depression, anxiety, post-traumatic stress disorder, migraine, and cluster headaches. LSD and psilocybin exert their psychedelic effects via activation of the 5-hydroxytryptamine 2A receptor (HTR2A). Here we provide a suite of engineered mice useful for clarifying the role of HTR2A and HTR2A-expressing neurons in psychedelic drug actions. We first generated Htr2a -EGFP-CT-IRES-CreERT2 mice (CT:C-terminus) to independently identify both HTR2A-EGFP-CT receptors and HTR2A-containing cells thereby providing a detailed anatomical map of HTR2A and identifying cell types that express HTR2A. We also generated a humanized Htr2a mouse line and an additional constitutive Htr2A -Cre mouse line. Psychedelics induced a variety of known behavioral changes in our mice validating their utility for behavioral studies. Finally, electrophysiology studies revealed that extracellular 5-HT elicited a HTR2A-mediated robust increase in firing of genetically-identified pyramidal neurons--consistent with a plasma membrane localization and mode of action. These mouse lines represent invaluable tools for elucidating the molecular, cellular, pharmacological, physiological, behavioral, and other actions of psychedelic drugs in vivo .
Serotonin (5-HT) modulates mood, perception, cognition, and a broad range of physiological functions through a family of 14 receptor subtypes, nearly all of which are G protein-coupled receptors. Of these, the 5-HT2A receptor (HTR2A) has attracted particular attention as the primary molecular target for classical psychedelic drugs — including LSD, psilocybin, DMT, and mescaline — as well as a high-affinity binding site for most atypical antipsychotics and many antidepressants. Despite decades of research, investigation of HTR2A function in psychedelic drug action has been hindered by the lack of validated genetic tools, with available reporter mouse lines showing inconsistent or inaccurate expression patterns relative to established autoradiographic and in situ hybridisation data. The present study aimed to develop and validate a suite of genetically engineered mice — collectively termed AMIS — to provide open-source platforms for interrogating the species specificity, cellular signalling, neural circuitry, and behavioural actions of psychedelic drugs. Using CRISPR/Cas9 genome editing, the researchers generated novel HTR2A-EGFP knock-in and HTR2A-EGFP-CreERT2 lines that faithfully recapitulate the endogenous expression and function of the receptor, enabling cell-type-specific circuit mapping and in vivo pharmacological interrogation with unprecedented precision.
Papers cited by this study that are also in Blossom
Gonza ´lez-Maeso, J., Weisstaub, N. V., Zhou, M. et al. · Neuron (2007)
Halberstadt, A. L., Geyer, M. A. · Neuropharmacology (2011)
Kim, K., Che, T., Panova, O. et al. · Cell (2020)
Kwan, A. C., Olson, D. E., Preller, K. H. et al. · Nature Medicine (2022)
CRISPR/Cas9 genome editing was used to insert a cDNA encoding EGFP into the C-terminus of the Htr2a locus on chromosome 14, after residue 452 of the receptor — a site previously validated as not affecting expression, function, targeting, or trafficking. The resulting Htr2a-EGFP-CT knock-in line was further modified to include an IRES-CreERT2 sequence to drive tamoxifen-inducible Cre recombinase expression under the endogenous Htr2a promoter, creating the Htr2a-EGFP-CT-IRES-CreERT2 line. Additional reporter lines and humanised HTR2A mice were generated using analogous strategies. Brain-wide distribution of HTR2A-EGFP-CT was characterised through immunohistochemistry, in situ hybridisation, whole-brain clearing with lightsheet microscopy, AChE staining, and saturation radioligand binding assays, and compared systematically with GENSAT reporter lines. Biochemical validation of receptor expression and downregulation capacity was performed via Western blot after pharmacological treatment. Designer receptor exclusively activated by designer drugs (DREADD) constructs were delivered via AAV to HTR2A-Cre mice to enable circuit-level manipulation. Behavioural pharmacology was assessed in both Htr2a-EGFP-CreERT2 and wild-type C57BL/6J mice following administration of LSD, DOI, and psilocin versus vehicle. Open field testing measured locomotion, rearing, and stereotypy over 30-minute blocks. Head-twitch responses (HTRs), grooming duration, and retrograde walking were quantified as psychedelic-specific behavioural proxies. Prepulse inhibition (PPI) of the acoustic startle reflex was assessed to examine sensorimotor gating. Whole-cell patch-clamp electrophysiology was performed in HTR2A-positive layer 5A pyramidal neurons of the medial prefrontal cortex.
The Htr2a-EGFP-CT knock-in mice showed a brain-wide HTR2A distribution closely consistent with published autoradiographic data. High expression was observed throughout deep cortical layers, the claustrum, striatum (in a striosomal pattern), ventral hippocampus (CA3), and lateral amygdala; expression was low or absent in the thalamus, midbrain, brainstem raphe nuclei, and cerebellum. In contrast, available GENSAT Htr2a-Cre lines showed aberrant patterns lacking striatal expression. Saturation binding assays and Western blot confirmed intact HTR2A function and normal downregulation capacity following agonist treatment. In behavioural experiments, basal locomotor activity was higher in Htr2a-EGFP-CreERT2 mice than in C57BL/6J mice. LSD, DOI, and psilocin all induced significant HTRs above vehicle levels in both genotypes; DOI was more potent in wild-type C57BL/6J mice for HTRs (p < 0.001), whilst LSD was more efficacious than psilocin in Htr2a-EGFP-CreERT2 mice. Grooming was broadly reduced by psilocin in both genotypes relative to vehicle; LSD and DOI stimulated retrograde walking to a greater extent than psilocin or vehicle. All three psychedelics disrupted PPI in both genotypes; however, PPI disruption was greater in C57BL/6J than in Htr2a-EGFP-CreERT2 animals, revealing genotype-dependent modulation of sensorimotor gating.
The genetically engineered mice described in this study provide validated tools for interrogating HTR2A expression and the behavioural and circuit-level actions of psychedelic drugs. The distribution of HTR2A-EGFP-CT was consistent with previous receptor autoradiographic and in situ hybridisation work, confirming high expression in cortical deep layers, claustrum, and the striatal striosomal compartment — regions implicated in psychedelic drug action. The discrepancy between the new knock-in lines and existing GENSAT reporter mice underscores the importance of rigorous validation before deploying genetic tools for circuit studies. Genotypic differences in behavioural responses to psychedelics between Htr2a-EGFP-CreERT2 and wild-type mice — particularly in HTR magnitude and PPI disruption — likely reflect differences in HTR2A expression levels resulting from the knock-in modification, and highlight the receptor's quantitative importance in mediating these effects. The CreERT2 platform enables tamoxifen-inducible genetic access to HTR2A-expressing neurons for targeted circuit manipulation using Cre-dependent reporters and actuators. Combined with humanised HTR2A mice that carry a single amino acid difference affecting psychedelic binding, and DREADD-based cell-type-specific activation tools, the AMIS suite provides a comprehensive open-source resource for the psychedelic neuroscience community. The researchers note that the suite enables direct investigation of species specificity in psychedelic pharmacology, a longstanding challenge in translating preclinical findings to the clinic.
Serotonin (5-hydroxytryptamine; 5-HT) is a biogenic amine neurotransmitter essential for modulating mood, perception, cognition, pain, feeding, and a variety of other functions in the central nervous system and periphery. To mediate these actions a large family of 14 distinct receptors that are heterogeneously distributed throughout the body have evolved. Except for the 5-HT3 receptor, all mammalian 5-HT receptors are G-protein-coupled receptors. Although the cell bodies of 5-HTproducing neurons are exclusively localized to brainstem sites, they project widely throughout the brain and spinal cord to regulate broad neural targets through 5-HT receptors. Of the various 5-HT receptors, the 5-hydroxytryptamine 2A receptor (HTR2A) has recently received the most attention, in part due to the apparent involvement of the HTR2A in the pathogenesis of several psychiatric disorders, including schizophrenia, depression, Parkinson's Disease psychosis, and anxiety disorders. Moreover, the HTR2A has been the focus of intense interest because this receptor is the primary molecular target for psychedelic drugs, including lysergic acid diethylamide (LSD), psilocybin, N,N'-dimethyltryptamine, and mescaline. The HTR2A also has high affinity for most second generation (atypical) antipsychotic drugs, as well as many typical and atypical antidepressants. At the cellular level, HTR2A agonists promote Gαq protein couplingand β-arrestin recruitment, leading to receptor internalizationand the activation of many downstream signaling pathways. The HTR2A was first identified by radioligand binding in 1978, was characterized as a membrane protein in 1985, and its encoding gene, Htr2a, was cloned in 1988. The structure of Htr2a was characterized further by the Chen and Shih groups for human HTR2A and Toth group for murine Htr2a, who identified various introns, exons, and promoters. The sequence of the HTR2A was generally conserved between species with the rodent Htr2a and human HTR2A genes sharing 91.5% deduced amino acid identity with a single amino acid difference in the binding pocket that alters the binding properties of many psychedelicand non-psychedelic 5-HT2A agonists. Initial receptor autoradiographicand in situ hybridization studiesidentified a heterogeneous distribution of HTR2A in the rat and human brains with high expression in the deep layers of the human and rodent cerebral cortex. Subsequent studies with HTR2A antibodiessuggested the HTR2A was localized primarily to pyramidal neurons with expression in somas, apical dendrites, and post-synaptic densitiesas well as occasional interneurons. Moreover, the HTR2A was expressed in subcortical regions, particularly in telencephalic sites. However, immunohistochemical studies of HTR2A location were conflicting and inconsistent (for example, seepresumably because almost none of the HTR2A antibodies used have been rigorously characterized. Only one antibody--which was suboptimal for routine studies--has been validated in wild-typeand Htr2a knock-out mice. Autoradiographic studies yielded somewhat more consistent findings, but most suffered from relatively low resolution as well as specificity issues related to radioligands. Similarly, in-situ hybridization revealed only mRNA levels, which do not necessarily correspond with protein expression. Thus, there is a pressing need for tools to precisely study HTR2A protein distribution and function. Here we provide as a resource to the scientific community, a suite of Htr2a reporter mice generated by CRISPR-mediated recombination at the endogenous Htr2a locus. We show this resource facilitates the identification and targeting of the HTR2A in vivo and provides potential opportunities for modulating the activity of HTR2A-expressing neurons with a variety of chemogeneticand optogeneticactuators. We further demonstrate the utility of these mice for electrophysiological, biochemical, and genetic studies of HTR2A-identified neurons. Given the essential role of the HTR2A in mediating the action of all known psychedelic and many non-psychedelic therapeutics, this resource will be of significant value to the scientific research community.
Prior to initiating our studies, we compared available Htr2a reporter mice that were developed as part of the GENSAT initiative. Two of these mouse strains, Tg(Htr2a-Cre)KM207Gsat/Mmucd and Tg(Htr2a-Cre)KM208Gsat/Mmucd, have reported distributions of GFP following Cre-mediated recombination with reporter mice (). As shown at similar coronal sections in Supplementary Figthese mice display distinct patterns of GFP expression. However, neither corresponds to the reported patterns of endogenous receptor binding, or for Htr2a mRNAor protein expression in rodents) (see. We note an absence of expression in the striatum in Tg(Htr2a-Cre)KM207Gsat/Mmucd and Tg(Htr2a-Cre)KM208Gsat/Mmucd miceeven though many prior studies have demonstrated significant HTR2A expression in this brain region (for example, see). In addition, recent mapping studies by the Allen Institute for Brain Science disclosed abundant Htr2a mRNA expression in murine cortex and in a patch-like distribution in striatum (; Suppl. Fig S1Cand), consistent with the observation that HTR2A is localized to striosomes in rodentbrains (FigS1D) and human. EGFP-CreERT2 ) mice were generated using CRISPR/Cas9 genome editing of exon 3 of Htr2a locus on chromosome 14 and a construct spanning nucleotides 74640840-74709494 to insert EGFP-CT followed by an IRES (internal ribosome entry site) and an estrogen-responsive Cre-recombinase (CreERT2). The mouse line has two modifications. First, a cDNA encoding EGFP was inserted into the C-terminus (CT) of the receptor, flanked on both ends by a serineand glycine-rich linker. The insertion site of EGFP was after residue 452, yielding a construct we term Htr2a-EGFP-CT. Parenthetically, this insertion site was identified based on our previous studiesand it was positioned between the 8 th helix and the C-terminal PDZ binding motif of the HTR2A. This location has been demonstrated not to affect expression, function, targeting, or trafficking of the receptor in neurons or transfected HEK293 cells. Second, the IRES-CreERT2 sequence directly followed the coding sequence, and it preceded the 3' UTR of Htr2a. The expression of CreERT2 recombinase under control of the endogenous Htr2a promoter provides a genetic manipulation platform to leverage Cre-dependent reporter mice or AAV viruses for specifically expressing reporter or actuator proteins in HTR2A containing cells.
To characterize these mice, we first examined whether EGFP cDNA inserted into exon3 of Htr2a genetic locus was expressed in a manner similar to the endogenous receptor. Here, in situ hybridization was used to detect the mouse Htr2a and Egfp transcripts in C57BL/6J (C57, Htr2a +/+ ) and Htr2a-EGFP-CT-IRES-CreERT2 mice. As shown in Figure, Egfp mRNA was expressed in the same cells as Htr2A mRNA in knock-in mice, but not in C57 animals. Next, we examined expression patterns/levels of HTR2A-EGFP-CT with GFP antibodies and immunofluorescent microscopy in fixed brain slices (Fig 2 ; Table) and by light-sheet microscopy in cleared brains (Movie S1-S4). Both strategies revealed similar results in corresponding coronal sections. A comparison of the expression patterns between HTR2A-EGFP-CTand the previously determined pattern of 3 H-M100907-labelled HTR2Arevealed excellent correspondence with expression of the receptor in our Htr2a EGFP-CreERT2 mice. A summary analysis of this receptor expression pattern in different brain regions follows in the next section, whereas a detailed description is found in Supplement #1.
The EGFP fluorophore property was used to locate HTR2A-EGFP-CT fusion proteins for the whole brain mapping. The HTR2A-EGFP-CT receptor was densely expressed in several forebrain areas, while in more caudal brain regions, receptor expression was restricted to only a few areas, most of which had relatively low expression. In neocortical areas, there was a dense band of HTR2A-EGFP-CT receptors in the most superficial aspect of layer 5a (L5a). Scattered HTR2A-EGFP-CT receptors were also present in L6 and supra-granular lamina with the exception of layer 1, where the apical tufts of HTR2A-positive pyramidal cells could be clearly visualized. In the frontal pole (FP), HTR2A-EGFP-CT in L5a formed a distinct continuous band that encircled the entire hemisphere, running parallel to the pial surface ( The densely-labeled L5a HTR2A-EGFP-CT cortical band was primarily comprised of pyramidal neurons with characteristic apical dendrites extending across L2/3 to enter and terminate in L1 as apical tufts (Figure). The dense HTR2A-EGFP-CT immunoreactivity that filled the somatodendritic compartments initially suggested cells in this cortical band were located in both superficial L5a and L4. However, the cell bodies of these HTR2A-EGFP-CT signals were predominantly found in L5a, with a characteristic pyramidal morphology rather than the small (stellate) cells of L4. Moreover, the cortical band of the HTR2A-EGFP-CT signal was present in the prelimbic cortex, which lacks a defined L4. We also stained sections for parvalbumin, a marker of the largest subset of cortical interneurons. Only a minority of neurons in the HTR2A-EGFP-CT cortical band expressed parvalbumin immunoreactivity (Figure), consistent with both pyramidal neurons and occasional interneurons expressing HTR2A. Early studies of the localization of HTR2A using autoradiographic methods concluded that the pyramidal cells that form a strongly-labeled band of cells across the cortex were in layer 4and layer 5However, these studies suffered from a lack of resolution. We therefore defined the laminar position of HTR2A-expresssing pyramidal cells by comparing HTR2A-EGFP-CT and well-characterized cortical laminar markers. These markers included N-terminal EFhand calcium binding protein 1 (NECAB1), which is primarily localized to L5a, and the L5b/L6 marker COUP-TF-interacting protein 2 (CTIP2). NECAB1 staining overlaped with the major band of cells expressing HTR2A-EGFP-CT. In contrast, the laminar distributions of CTIP2-and HTR2A-EGFP-CT expression did not overlap. A validated HTR2A antibodywas also used to determine if the expression pattern of the HTR2A protein corresponded with the GFPstained HTR2A-EGFP-CT fusion protein; both antibodies displayed similar staining patterns. Therefore, the prominent layer expressing dense HTR2A-EGFP-CTpositive pyramidal cells is L5a and not L4. HTR2A-EGFP-CT in the telencephalon was not restricted to the cortex but was also present in subcortical sites. In the dorsal striatum (caudatoputamen, CPu), HTR2A-EGFP-CT densely filled the striosomal compartments (CPu-s) (defined on the basis of mu opioid receptor (MOR) immunoreactivity). HTR2A-EGFP-CT was co-localized with MOR in striosomes, including the subcallosal stria (Fig1H). HTR2A-EGFP-CT was present but much less dense in the matrix (CPu-m) compartment (
In mice and rats, acute administration of psychedelic drugs produces a stereotypical head twitch response, disruption of prepulse inhibition (PPI), and variable changes in motor behaviors. Accordingly, we next evaluated these responses to vehicle, 0.3 mg/kg LSD, 1 mg/kg 2,5-dimethoxy-4-iodoamphetamine (DOI), and 1 mg/kg psilocin (i.p.) in C57 and Htr2a EGFP-CreERT2 mice to examine whether HTR2A-EGFP-CT function was similar to endogenous HTR2A.
In the open field baseline activities were collected in 5-min blocks over 30 min and post-injection responses were followed over the next 90 min. Locomotor activities were increased overall in both genotypes by DOI relative to vehicle (p-values≤0.031) (Suppl. Overall responses to the vehicle and psilocin were not distinguished from each other, while both LSD and DOI augmented stereotypical behaviors over these two groups (pvalues!"#"$%&#'' Together, these results show that basal motor activities were higher in the Htr2a EGFP-CreERT2 than in C57 mice and that DOI was more potent in stimulating locomotion in the knock-in animals. Following vehicle or psychedelic administration, these genotype effects persisted in locomotion and rearing with ANCOVA, whereas DOI in locomotion and LSD and DOI in rearing and stereotypes were more potent in stimulating activities compared to psilocin and the vehicle control.
LSD, DOI, and psilocin stimulated more head twitch responses (HTRs) in the C57 and Htr2a EGFP-CreERT2 mice relative to the vehicle controls (p-values!"#""(&)' *+,-,' -,./01.,.' *,-,' 2,-3' 40*' 5678' (9&#' ' 9:;' *<.' =0-,' /0>,1>' 71' .>7=?4<>718' head twitch responses'71'@AB'>+<1'71'Htr2a EGFP-CreERT2 mice (p<0.001), with a strong trend for psilocin (p=0.053). No genotype differences were evident between the vehicle and LSD treatments. Within C57 mice, all psychedelics stimulated more head twitch responses than vehicle (p-values<0.001). Additionally, DOI stimulated more head twitch responses than LSD or psilocin (p-values≤0.043), which were not statistically different from each other. ;1'Htr2a EGFP-CreERT2 animals, the numbers of head twitch responses were lower with vehicle than all hallucinogen-treated groups (p-values≤0.003), and LSD was more efficacious than psilocin (p=0.026). Thus, both genotype and treatment effects in head twitch responses were observed. Time spent grooming was also examined. Grooming was more prolonged in the C57 mice given LSD and DOI relative to that in the Htr2a EGFP-CreERT2 animals (p-values<0.001). Within C57 subjects, both LSD and DOI stimulated grooming for longer periods of time than those observed in the psilocin-and vehicle-treated mice (pvalues!"#""C&)'*+7D+'*,-,'10>'E7FF,-,1>'F-0=',<D+'0>+,-#''G3'D0=/<-7.01)'71'Htr2a EGFP- CreERT2 animals groom time was reduced with psilocin relative to the LSD and vehicle groups (p-values≤0.008). In summary, grooming is largely reduced by psilocin in both genotypes relative to vehicle, whereas genotype differences are evident with LSD and DOI. No genotype distinctions were found for retrograde walking. Here both LSD and DOI stimulated this response to greater extents than in the psilocin-and vehicletreated mice (p-values<0.001).
Genotypic differences were evident where PPI was lower overall in C57 than in Htr2a EGFP- CreERT2 mice (p<0.001). Moreover, within genotypes all three hallucinogens disrupted PPI relative to vehicle (p-values<"#""C&#''G0>+'1?44'<1E'.><->4,'<D>727>7,.'*,-,' <1<43H,E#' :2,-<44' 1?44' <D>727>3' *<.' +78+,-' 71' Htr2a EGFP-CreERT2 than in C57 animals (p-values<0.001) -5I?//4#'678'I(J&#''J1'<..,..=,1>'0F'>-,<>=,1>',FF,D>.'-,2,<4,E'02,-<44' 1?44'<D>727>7,.'*,-,'40*,-'71'>+,'2,+7D4,'>+<1'71'>+,'KI9'<1E'9:;'8-0?/.'5p-values<0.001). In addition, null activity was higher in DOI-than in LSD-and psilocin-treated mice (p-values<0.001). Genotype effects were observed also for startle activities. Here responses to the vehicle, LSD, and psilocin were reduced in Htr2a EGFP-CreERT2 , compared to the C57 animals (p-values<0.001). Within C57 mice startle activities were higher overall in the psilocin than in the vehicle or LSD groups (p-values!"#"LL&. In Htr2a EGFP-CreERT2 mice, startle to DOI was higher than in the vehicle and LSD animals (pvalues!"#"C"&#' ' @044,D>72,43)' >+,.,' F71E718.' 71E7D<>,' >+<>' MM;' *<.' E7.-?/>,E' 71' N0>+' 8,10>3/,.'*7>+',<D+'0F'>+,'/.3D+,E,47D.O'+0*,2,-)'MM;'*<.'=0-,'/,->?-N,E'71'@AB'>+<1' 71'>+,'Htr2a EGFP-CreERT2 mice.
[ 3 H]-ketanserin saturation binding assay was used to quantify the affinity (Kd) and number of HTR2A-EGFP-CT receptors (Bmax) in the whole cerebral cortex (Fig). C57 and Htr2a EGFP-CreERT2 mice showed no statistically significant differences in Kd values (p=0.35) although receptor expression in the Htr2a EGFP-CreERT2 mice was statistically enhanced compared to C57 controls (p=0.0005). Receptor mRNA levels via real-time qPCR were examined from the whole brain cortex in C57, as well as in heterozygous and homozygous Htr2a EGFP-CreERT2 mice (Fig 4B). Htr2A mRNA expression levels in Htr2a EGFP- CreERT2 mice were statistically enhanced in the heterozygous (p=0.0002) and in the homozygous Htr2a EGFP-CreERT2 mice (p<0.0001) relative to C57 controls. Increased transcript levels have been previously reported for other 3' manipulations of GPCRs. It is well known that LSD and other psychedelic drugs induce substantial HTR2A downregulation following repeated in vivo administration. Here, C57 and Htr2a EGFP-CreERT2 mice were treated with 0.5 mg/kg LSD (i.p.) for 5 consecutive days and 24 hr later the whole cerebral cortex was removed for receptor purification and western blot analyses as previously described. After 5-days of daily LSD administration, endogenous HTR2A in C57 and HTR2A-EGFP-CT fusion receptors from Htr2a EGFP-CreERT2 mice were downregulated to similar extents compared to the vehicle-treated group (p<0.0001). These results indicate that the EGFP insertion in the C-terminus of Htr2A does not affect in vivo receptor down-regulation response to LSD which is in agreement with prior studies using a similar strategy to investigate internalization and down-regulation in vitro.
The bi-cistronic design of Htr2a-EGFP-CT-IRES-CreERT2 produced a HTR2A-EGFP-CT fusion protein and CreERT2 recombinase. The expression of CreERT2 recombinase under Htr2a promoter provided a genetic manipulation platform via Credependent reporter (Ai9 mice here) for specifically expressing tdTomato reporter in HTR2A containing cells. HTR2A-tdTomato expressing neurons represent HTR2Apositive neurons, while the HTR2A-EGFP-CT fusion protein represents HTR2A receptor itself. Whole brain mapping of induced tdTomato expression (Suppl. Fig S4 ) was compared with HTR2A-EGFP-CT expression in slices and cleared whole brain. The result from whole brain cleared tissue showed HTR2A-EGFP-CT and tdTomato to have quite similar distributions in the 3D rendering of coronal, horizontal, and sagittal views, respectively (movie S1, S2, S3 and S4). For controls, the Htr2a EGFP-CreERT2/+ x Ai9 and C57x Ai9 mice were treated with vehicle (corn oil) or tamoxifen to determine whether tdTomato expression was specifically regulated by tamoxifen for driving CreERT2 recombinase. Only Htr2a EGFP-CreERT2/+ x Ai9 mice with tamoxifen treatment showed tdTomato expression, whereas none of the other treated groups showed this expression.
We leveraged the visualization properties of tdTomato-expressing neurons to conduct electrophysiological whole-cell recording in ex vivo brain sections, allowing us to study 5-HT-mediated responses in HTR2A-containing pyramidal neurons in the cortical L5a of the mPFC. We first characterized the mPFC subregion distributions of HTR2A-EGFP-CT and HTR2A-tdT in Htr2a EGFP-CreERT2/+ x Ai9 mice. The cytoarchitecture features by AchE staining revealed an AchE-weak zone in the ventral PL and IL. Nevertheless, HTR2A-EGFP-CT was robustly expressed in cortical L5a of the ACC and dorsal PL, but not in the ventral PL and IL; tdTomato-positive neurons also had similar subregion distributions (Fig). Thus, we specifically targeted tdTomato-expressing neurons in the ACC and dorsal PL (dPL) for electrophysiological recordings. Recent research has questioned whether 5-HT can activate HTR2A in vivogiven that the HTR2A protein has been shown to reside both in intracellular compartmentsand at postsynaptic sites. To test this hypothesis, we performed whole-cell current-clamp recordings of neuronal firing in visually-identified tdTomato-expressing neurons in L5a of the ACC or dPL. Focal application of 5-HT (10 µM) produced a biphasic response, with an immediate but transient reduction in firing followed by a sustained increase in firing frequency. Following the electrophysiological recordings, the recorded tdTomato neuron was filled with biocytin, which facilitated the verification of its anatomical location. The location of the biocytin-labeled tdTomato neuron was confirmed subsequently using a fluorescent streptavidin and co-staining with a GFP antibody for HTR2A-EGFP-CT labeling. This costaining revealed the soma of the recorded neuron resided in the GFP-positive layer, L5a within the ACC or dPL. Moreover, the biocytin filling of the entire neuron revealed its pyramidal neuron morphology. During the recording process, each recorded neuron was imaged under a fluorescence microscope to provide a visual representation of its location within the ACC or dPL. The schema depicting the location of individual recorded neurons is shown in Supplementary Figure.
We generated two additional Htr2a constitutive Cre mouse lines. The first one is the Htr2a-IRES-Cre mouse line (Htr2a Cre ) that was designed with aa native Htr2a coding sequence followed by IRES-Cre in the 3'UTR. The Htr2a genomic coding sequence was not modified in this mouse when inserting the IRES-Cre into the 3'UTR, in contrast to the already described Htr2a-EGFP-CT-IRES-CreERT2 mouse. The second line is a humanized line, Htr2a-A242S-EGFP-CT-IRES-Cre (Htr2a A242S-EGFP-Cre ), which was designed so that a single point mutation in alanine residue 242 was replaced by a serine from the human HTR2A and the EGFP was inserted into the C-terminus followed by the IRES-Cre sequence. This single point mutation has been shown previously to enhance the affinity and potency of a variety of psychedelic and non-psychedelic drugs to levels similar to those seen at the human receptor. Receptor distributions were visualized with an anti-HTR2A antibody for HTR2A in the Htr2a Cre mice (Suppl. The distribution of the HTR2A-containging cells was examined also through recombinant expression of tdTomato. As Cre-recombinase activity is driven at various stages of development, many cortical neurons displayed tdTomato fluorescence in a pattern quite distinct from that of endogenous HTR2A expression (Suppl. As a comparison, inducible Htr2a EGFP-CreERT2 mice were retro-orbitally injected with PHP.eB.FLEX.tdTomato to assess their tdTomato expression patterns. Interestingly, differences were observed in the tdTomato expressing patterns between the virallymediated approach and the Ai9 reporter mice in the inducible mice (Suppl. Figand). For example, the viral approach revealed trace levels of the tdTomato signal in the striosome (CPu-s) and olfactory tubercle (TUO), whereas the signal was high in the induseum griseum (IG) and septohippocampal nucleus (SH) (Suppl FigS7D1). In the cortex, virally-mediated tdTomato expressing neurons were found in the deeper cortical layers L5b and L6 (minor), while Ai9-mediated inducible tdTomato was located in L5a and L6--but not in L5b. Despite the similar distribution of HTR2A receptors observed in all three mouse lines, different Cre-dependent approaches (reporter viruses and reporter mice) displayed dissimilar expression patterns of HTR2A-containing cells due to distinct Cre recombination occurring timing and efficiencies (Cre vs. CreERT2), and viral infection efficiencies (resulting in incomplete cell infected) as well as various promoters of Cre-dependent reporters. Notably, when using Cre-dependent reporters as a readout investigates HTR2A-mediated effects, it will be important to validate whether reporter expression patterns correspond with HTR2A receptor distributions in the same brain region. As previously mentioned, the human HTR2A differs from the rodent version of the receptor by virtue of a single amino acid change in the binding pocket (S242A) which attenuates the activity of a variety of psychedelicsand non-psychedelic HTR2A agonists. Accordingly, we created mutant mice in which murine Ala242 was mutated to the human serine residue (i.e., A242S). Initial behavioral studies were conducted to compare head twitch responses and PPI performance between the C57 and Htr2a A242S-EGFP-Cre mice. In C57 animals, DOI elicited more robust head twitch responses than in the Htr2a A242S-EGFP-Cre mice (p<0.001), whereas the converse was observed with psilocin (p=0.026). In both genotypes, all psychedelics stimulated head twitch responses relative to the vehicle control (p-values<0.001), while the response to DOI was more pronounced than that to psilocin but only in the C57 mice (p=0.001). An examination of PPI revealed responses were higher in the Htr2a A242S-EGFP-Cre than in the C57 mice (p=0.009) and all psychedelics were disruptive of PPI (pP2<4?,.!"#""C&' 5I?//4#' 678' QG&#' ' ;1' <EE7>701)' 02,-<44' MM;' -,./01.,.' *,-,'+78+,-'>0'/.740D71'>+<1'9:;'5/R"#"LL&#'S,10>3/,',FF,D>.',=,-8,E'F0-'1?44'<D>727>7,.' 5/T"#""C&)'<1E'>-,<>=,1>',FF,D>.'*,-,',27E,1>'5I?//4#'678'Q@&#''@0=/<-,E'>0'2,+7D4,)'1?44' <D>727>7,.' *,-,' ,1+<1D,E' N3' <44' /.3D+,E,47D.' 5p-values<0.016). Startle activities were higher in C57 than Htr2a A242S-EGFP-Cre mice given the vehicle or psilocin (pP2<4?,.!"#"$A&' 5I?//4#'678'Q9&#'';1'@AB'<17=<4.)'.><->4,'*<.'+78+,-'>0'/.740D71'>+<1'>0'KI9'5p=0.004), whereas in Htr2a A242S-EGFP-Cre mice responses to DOI were enhanced over vehicle (p<0.001). Collectively, these results show that while overall responses to the different psychedelics are relatively uniform for PPI, there are some genotype differences in HTRs. Biochemical studies in Htr2a A242S-EGFP-Cre mice have found the HTR2A affinities as Kd values to be 2.38 ± 1.48nM and receptor expression levels as Bmax values to be 0.49 ± 0.2 pmole/mg using [ 3 H]-ketanserin radioligand binding assays. Given the welldocumented differences between human and rodent HTR2A psychedelic pharmacology, the A242S mouse line will likely prove useful for interrogating the action of psychedelic drugs in vivo. We also compared the distribution of HTR2A in neurons and those labelled with the Thy1-GFP reporter which previously had been used to quantify spine formation after psychedelic drug administration. For these studies, our Htr2a EGFP-CreERT2 x Ai9 mice were crossed with Thy1-GFP mice. As can be seen in the cortex (Suppl. Fig S9), Thy1-GFP neurons did not co-localize with the tdTomato-expressing neurons in L5a. Only rare Thy1-GFP-positive/HTR2A-tdTomatopositive neurons were observed in L5b/L6 (tdT/Thy1-GFP in L5b/L6 of the ACC/dPL cortex: 1.23 ±1% from 445 neurons of 3 animals). Given the lack of 1:1 co-localization with Thy1-GFP, our results imply that studies on psychedelic drug-induced plasticity using this reporter mouse may not faithfully quantify spine formation in HTR2A-expressing neurons. Thy1-GFP mice are useful, however, to determine the effects of psychedelics and other drugs on non-HTR2A-expressing neurons.
We also manipulated the activity of HTR2A-expressing neurons with a chemogenetic actuator (deschloroclozapine, DCZ) in Htr2a Cre mice with local viral injection of DREADD into the mPFC. Htr2a Cre/+ mice that expressed excitatory DREADD (DIO-HA-hM3Dq-IRES-mCitrine) displayed a significant depolarization response to bath application of DCZ, whereas HTR2A + neurons expressing DIO-eYFP did not show a physiological response. Consequently, DCZ was shown to specifically activate hM3Dq + /HTR2A + neurons. This illustrates how Htr2a Cre/+ mice prove to be valuable in chemogenetic studies for revealing the functions of specific cell populations expressing HTR2A.
Psychedelic drugs have emerged as transformative therapeutics for a large number of neuropsychiatric disordersand all classical psychedelics induce their typical responses in miceand humansvia HT2ARs. Despite decades of study, however, we lack key technologies to interrogate the role of the HTR2A in the actions of these drugs. Here we provide an open-source suite of validated genetically-engineered mouse models (AMIS,) which provides platforms to interrogate the species-specificity, cellular signaling, neural circuitry, and behavioral actions of psychedelic drugs. Previous studies have utilized distinct and orthogonal approaches to examine the distribution of HTR2A by receptor autoradiography and in situ hybridization, as well as by studying transgenic mice. The early autoradiography work in rats from the Palacios groupand othersshows a similar distribution of the HTR2A in the brain as the results with our engineered mouse models. HTR2A expression was high in the claustrum, olfactory tubercle, piriform cortex, and what was (erroneously) designated layer 4 of the cortex. Moderate levels were seen in the striatal complex and certain nuclei of the amygdala, but outside of the forebrain, in regions such as the thalamus, midbrain, and lower brainstem, HTR2A was not present or seen in low levels. The distribution of the HTR2A has been studied also in human brain using [ 3 H]ketanserin, [ 3 H]-mesulergine, [ 3 H]-LSD, and [ 3 H]-spiperone. Of these ligands, ketanserin is the most selective for the HTR2A and the expression patterns with this radioactive ligand more closely correspond with patterns we observed using our mouse Htr2a EGFP-CreERT2 line. The similarities were evident in comparisons of high expressing cortical areas; intermediate expression was seen in the claustrum and striatum as well as restricted areas of the amygdala, and very low or absent labeling in the thalamus, brainstem, and cerebellum. In our knock-in mice, HTR2A-EGFP-CT showed a similar trend. High expression levels were observed throughout the cortex, in the striatal complex, and claustrum. Other forebrain loci with high expression included CA3 in the ventral (temporal) hippocampus and the lateral nucleus of the amygdala. In more posterior areas high expression was observed in the medial mammillary nucleus and both the basis pontis and the inferior olive. HTR2A-EGFP-CT expression of was either absent or seen in trace levels in diencephalic (thalamus and hypothalamus) and midbrain areas, as well as the ponts and medulla. Notable was the lack of expression in the brain raphe nuclei, with the exception of low levels in the raphe obscurus and raphe magnus (see Table). The GENSAT resourceprovides transgenic mice for interrogating HTR2A circuits and function (see Htr2a-EGFP-
Htr2a-Cre-KM207 (FigS1A) and Htr2A-Cre-KM208 (FigS1B) ()) among others. Unfortunately, these lines do not faithfully recapitulate the endogenous distribution of the HTR2A as has been seen with other BAC-transgenic mice. Despite this limitation, some groups have been used for HTR2-related research, such as electrophysiological studies, retina-related studiesand circuitry-related research. The resultant data should be interpreted with caution since the distribution of HTR2A in the Htr2a BAC mouse lines are distinct from that of the endogenous receptor based on studies using receptor autoradiography, immunohistochemistry, in situ hybridizationand our mouse lines. We also observed Htr2a transcript distribution, which showed some modest differences compared with HTR2A-EGFP-CT protein expression (Suppl. Fig S10A ). Htr2a transcripts displayed cortical distribution from L2/3, L5 to L6, while HTR2A-EGFP-CT was mainly expressed in L1, L5a and L6. These results suggest some level of dissociation between the mRNA and protein levels of the HTR2A in specific cortical layers as should be expected since the HTR2A protein is localized primarily to processes rather than cell bodies. As a follow-up for mRNA expression, we examined whether Egfp-tagged Htr2a transcripts were normally expressed in Lamp5-positive neurons (a marker for L2/3 and L5b), parvalbumin inhibitory interneurons, and Cplx3-positive neurons (a marker for L6b). We found Htr2a-Egfp transcripts were co-localized with Lamp5 in L2/3 and L5b, as well as with the interneuron marker parvalbumin and with Cplx3 in L6b (Suppl. Fig). These results are consistent with the Allen Institute for Brain Science Brain Map for the distribution of Htr2a mRNA in cortical L2/3, L5, and L6 (). The data we have obtained on the localization of HTR2A-EGFP-CT may offer some clues as to the sites of action of psychedelic drugs. Doss et alhave recently reviewed and evaluated the three major current hypotheses on brain circuits that underlie the psychedelic actions of psychedelic drugs. All three involve a large expanse of cortical areas, but otherwise differ. One of these models is a cortico-striato-fugal circuit that returns to (and goes beyond) the originating cortical site. This Cortico-Striatal-Thalamo-Cortical (CSTC) model posits a key role for thalamic projections from pallidal sites that receive striatal inputs, with prominent roles for the thalamic reticular nucleus and the mediodorsal nucleus (MD). In an immunohistochemical studycommented on the presence of low but detectable levels of HTR2A in the thalamic reticular nucleus, the antibody used was not characterized. Using in situ hybridization histochemistry,observed very low levels of the Htr2a transcript in the reticular or mediodorsal thalamic nuclei. Almost all other studies, including the autoradiographic studies using radiolabeled The mouse lines we have developed as a resource for the scientific community should be of considerable utility in testing models of brain circuits subserving psychedelic action. In particular, our data and those of other investigators point to the REBUS and CCC models as being the most plausible. We also examined several behaviors that have been linked to psychedelic drug actions in rodents. For instance, drug-induced changes in locomotor activity have been reported with LSD, DOI, and psilocin. The HTR2A has been implicated in the modulation of locomotor activity in rodentsand because this receptor is expressed in the striatum, HTR2A agonists could modulate the dopaminergic and serotonergic systems to affect motor activity. Notably, the selective HTR2A antagonist M100907 blocks these effects. Due to their robust polypharmacology, psychedelics like LSD potently activate several dopamine receptorsand this dopaminergic activity could contribute to these locomotor responses. The head twitch response is a validated animal model for identifying psychedelic drugs in rodentsand it is mediated by the activation of HTR2A in the mPFC. Previous studies have demonstrated that the LSD-, DOI-, DMT-, and psilocybin-induced head twitch responses in rodents are blocked by HTR2A antagonistsand by genetic deletion of Htr2a in mice. Another proxy for the actions of psychedelic drugs is PPI which is abnormal in schizophrenic patientsand is disrupted in both humansand rodentsby psychedelic, but not by non-psychedelic HTR2A agonists. In behavioral tests of our HTR2A reporter mice, we used psychedelic drugs as positive controls because they induce hallucinations in humans and are reported to have potential therapeutic benefits. We observed LSD and DOI stimulated motor activity, head twitches, retrograde walking, and produced disruption of PPI; psilocin increased HTR and disrupted PPI. Taken together, our results indicate that our engineered mouse models display typical responses to psychedelic drugs using a variety of behavioral readouts. In electrophysiological studies, 5-HT acted on genetically-identified L5a pyramidal neurons in the mPFC, producing a biphasic effect wherein a brief inhibition of firing (1 min) was immediately followed by a prolonged HTR2A-mediated excitatory effect in the late phase (>3 min). The HTR1A antagonist partially reversed the 5-HT mediated inhibitory effect on firing, while the HTR2A antagonist completely blocked 5-HT-mediated increase in firing. This result is consistent with previous findings showing opposite effects of HTR1A (hyperpolarization) and HTR2A (depolarization) in single cell activities. Thus, the HTR2A plays a major role in 5-HT-induced increases in neuronal activity in HTR2A + /L5a pyramidal neurons in the mPFC. These findings are important because of a recent report suggesting that many HTR2A agonists may exert their actions via intracellular rather than plasma membrane receptors. As we recently reported, psychedelic drugs increase the firing of genetically-identified HTR2A neurons in a manner similar to that reported here for 5-HT. As 5-HT is extracellularly applied, our results are consistent with the hypothesis that the psychedelic actions of HTR2A agonists are mediated by direct actions of pyramidal neuron firing via surface HTR2A receptors. With respect to the HTR2A mouse lines we have generated, they can be used as an integrated toolset for studying psychedelic-mediated mechanisms. In the present study, we demonstrated the whole brain mapping of HTR2A All behavioral experiments were conducted with an approved protocol from the Duke University IACUC. All experiments were performed under relevant regulations and ARRIVE guidelines. Prior to behavioral testing, the C57BL/6J mice had been screened in the zero maze for anxiety-like behavior, open field for locomotor activity, and prepulse inhibition for sensorimotor gating and mice that responded "normally" on all three tests were used to further backcross the knock-in mice onto a C57BL/6J genetic background. These C57BL/6J mice were bred for 2 generations with the knock-in mice to render them most similar in behavioral phenotype to their native inbred animals.
We generated two mouse HTR2A lines, Htr2a-IRES-Cre and the Htr2a-EGFP-CT-IRES-Cre ERT2 . We then produced an additional reporter line that humanized the HTR2A in its orthosteric binding site. The Htr2a-A242S-EGFP-CT-IRES-Cre line humanizes the ligand binding pocket by introducing the A242S as a single point mutation. We used the GRCm38/mm10 genome assembly for all annotations listed below. The Htr2a gene is on chromosome 14 and spans nucleotides 74640840-74709494. EGFP was inserted into the C terminal tail of the receptor, flanked on both ends by a serine-and glycine-rich linker. This insertion occurs after residue 452, corresponding to a DNA insertion site after chr14:74706337. This location is ideal in that it is distal to the final helix of the receptor so as not to disturb protein folding and also distal to the PDZ domain so as to not disturb protein trafficking. The IRES-Cre or IRES-CreERT2 sequence was inserted directly following the coding sequence and preceding the 3' UTR, corresponding to a genomic insertion after nucleotide chr14:74706397. The A242S point mutation was introduced by changing GCA at nucleotides 74705705-74705707 to TCC.
For the production of recombinant Cas9 protein, a human codon optimized FLAG-Cas9 C57BL/6J zygotes were microinjected with 400 nM Cas9 protein, 25 ng/μl of each guide RNA, and 20 ng/μl donor vector. In some cases, in vitro transcribed mRNA encoding codon-optimized Flp recombinase was included in the microinjection mix at 50 ng/μl to reduce tandem integration events by recombining the single FRT sites in tandem copies of the donor vector. To prepare the microinjection mix, guide RNAs were diluted in microinjection buffer, heated to 95°C for 3 min, and placed on ice prior to addition of Cas9 protein. The mixture was then incubated at 37°C for 5 min and placed on ice, after which the donor vector was added, and the mixture was held on ice prior to pronuclear microinjection. Microinjected embryos were implanted in recipient pseudopregnant B6D2F1/J females (#100006; Jackson Labs). Resulting pups were screened by PCR and sequencing for the presence of the correct insertion allele. One or more founders with the correct allele were mated to inbred C57BL/6J animals to transmit the modified allele through the germline. Offspring from a single founder line were selected for additional breeding to maintain and characterize the line.
The drugs used in the behavioral experiments were (±)-2,5-dimethoxy-4iodoamphetamine hydrochloride (DOI; Bio-Techne Corp., Minneapolis, MN), (+)-lysergic acid diethylamide-(+)-tartrate (LSD), and psilocin (NIDA Drug Supply Program, Bethesda, MD). The vehicle was composed of N,N-dimethyllacetamide (final volume 0.5%; Sigma-Aldrich) that was brought to volume with 5% 2-hydroxypropoyl-β-cyclodextrin (Sigma-Aldrich) in water (Mediatech Inc., Manassas, VA). All drugs were administered (i.p.) in a 5 ml/kg volume. For electrophysiological experiments, serotonin hydrochloride and M100907 were purchased from Sigma-Aldrich and RS102221 was purchased from Tocris Bioscience (Ellisville, MO). Stock solutions were prepared in ultra-pure water or DMSO, stored at -20°C and diluted to working concentration in aCSF on the day of testing.
The RNAscope experiment followed the ACDBio protocol (RNAscope Multiplex Fluorescent Reagent Kit v2 assay) with slight modifications. Fresh brains from Htr2a +/+ and Htr2a EGFP-CreERT2 mice were dissected and immediately embedding into O.C.T specimen matrix (Tissue-Tek® O.C.T Compound, Sakura ®Finetek). Brain sections at 18 µm thickness were cut using a cryostat and directly mounted onto slides (Fisherbrand TM Superfrost TM Plus microscope slides) followed by a 1 hr post-fixation with 4% paraformaldehyde (PFA) in PBS at 4°C. Brain sections were passed through a serial gradient ethanol dehydration step, followed by H2O2 treatment to quench endogenous peroxidase activity. The sections were next subjected to protease 3 digestion for 20 min. Mouse Htr2a (#401291-C3) and Egfp (#400281) probes were used for hybridization over 2 hr at 40°C in a humidity-controlled oven (HybEZII, ACDBio) and then the signal was amplified using Opal Dye570 for the Egfp probe and Opal Dye690 for the Htr2a probe. Pvalb (#421931-C2), Lamp5 (#451071), and Cplx3 (#467821-C3) were used as probes for interneuronal and cortical layer markers. Slides were counterstained with DAPI and mounted. Images were collected under an Olympus VS120 virtual slide microscope and Olympus FV3000RS Confocal Microscope (Olympus, Tokyo, Japan).
C57, Htr2a EGFP-CreERT2 , Htr2a Cre and Htr2a A242S-EGFP-Cre mice were crossed with the Credependent reporter Ai9 strain. Mice were injected (i.p.) with vehicle (corn oil, Sigma) or 100 mg/kg tamoxifen (Sigma) in a 5 ml/kg volume from p39 to p42. At 14 days post injection (14 dpi), mice were euthanized at p56, and brains were dissected for further HTR2A expression surveys. In a separate experiment, C57, Htr2a EGFP-CreERT2 , Htr2a Cre and Htr2a A242S-EGFP-Cre mice were retro-orbitally injected with AAV-PHP.eB-FLEX-tdTomato. Brains were collected later for analysis. The pAAV-FLEX-tdTomato was a gift from Edward Boyden (Addgene plasmid #28306, Addgene, Watertown, MA). Mice were intracardially perfused with 4% PFA in PBS. Brains were harvested, post-fixed in 4% PFA/PBS overnight, and dehydrated in 30% sucrose/PBS until sinking. Brains were sectioned by cryostat at 40 µm. The free-floating brain sections were washed 3 times with 0.1% TX-100/PBS prior to a 1 hr incubation in blocking buffer (5% normal donkey serum in 0.4%TX-100/PBS) and then incubated overnight at 4°C with anti-goat GFP (1:1000, Rockland, #600-101-215), anti-rabbit NECAB1 (1:1000, Sigma, # HPA023629), anti-Parvalbumin (1:4000, Swant, #235), and anti-rabbit mu opioid receptor (MOR) (1:4000, a gift from Dr. Lee-Yuan Liu-Chen, Temple University;)). The anti-rabbit HTR2A antibody (1:250, Neuromics, #RA24288) and anti-rat CTIP2 (1:1000, Abcam, #ab18465) primary antibodies were incubated at room temperature (RT) overnight. Brain sections were washed with 0.1% TX-100/PBS 3 times and then incubated with secondary antibodies (donkey anti-goat-Alexa Fluor® 488 for GFP primary antibody; anti-rabbit-Alexa Fluor® 647 for 5-HT2A, NECAB1, and MOR primary antibodies; and anti-rat-Alexa Fluor® 647 for CTIP2 at 1:1000 dilution; Jackson Immunoresearch, West Grove, PA). Tissue sections were imaged on an Olympus VS120 virtual slide microscope or Olympus FV3000RS Confocal Microscope (Olympus, Tokyo, Japan).
The C57 or Htr2a EGFP-CreERT2 x Ai9 mice (treated with tamoxifen) were used in this study. Mouse brains were processed with a newly optimized protocol (U.Clear,Cosmo Bio USA) with the refractive index adjusted to 1.53 using 2,2'-thiodiethanol. Samples were stored at -20 °C until acquisition. The cleared brain samples were imaged horizontally with tiling using the LifeCanvas SmartSPIM lightsheet microscope. 561/647 nm lasers were used for GFP/RFP/IHC imaging with the 3.6×/0.2 detection lens. Lightsheet illumination was focused with a NA 0.2 lens and axially scanned with an electrically-tunable lens coupled to a camera (Hamamatsu Orca Back-Thin Fusion) in slit mode. The camera exposure was set at fast mode (2 ms) with 16b imaging. The X/Y sampling rate was 1.866 μm and tha Z step was set at 2 μm.
The method for AchE staining method was as reportedwith modifications as follows. Brain sections (40 μm) were cut and then mounted onto slides. The slides were incubated for 6 hr with solution A [0.0072% ethopropazine, 0.075% glycine, 0.05% cupric sulfate, 0.12% acetyl thiocholine iodide, 0.68% sodium acetate, (pH 5)] for enzymatic reactions. Slides were washed with ddH2O 3 times for 5 min and then were developed in solution B (0.77% sodium sulfide, pH 7.8) for 30 min. Next, slides were washed 3 times for 5 min and were exposed to solution C (1% silver nitrate) for 10 min in the dark for silver intensification and then rinsed for 3 times over 5 min. The slides were air-dried overnight and then were cleared in two changes of xylene followed by Permount mounting media.
The whole brain cortices were collected from C57, Htr2a EGFP-CreERT2 , and Htr2a A242S-EGFP- Cre mice. Cortices was homogenized by a polytron homogenizer (BioSpec Products, Inc., Bartlesville, OK) on ice in 50 mM Tris buffer (pH7.4) with 0.1 mM PMSF and 1 mM EDTA. The homogenate was centrifuged at 1,000 x g for 10 min at 4°C. The supernatant was reserved for further centrifugation at 31,000 x g for 15 min at 4°C; pellets were collected as crude membranes. The crude membranes were suspended in 50 mM Tris (pH7.4) and 320 mM sucrose and were passed through a 26.5 G needle 5 times and then stored at -80°C until use. The saturation binding assay was conducted using an 8-point dosedependent concentration of [
The whole brain cortices from C57, Htr2a EGFP-CreERT2/+ , and Htr2a EGFP-CreERT2/EGFP-CreERT2 mice were dissected for total RNA purification. Cortices were homogenized with 1 ml Trizol (ThermoFisher) followed by chloroform (0.2 ml) extraction. The mixture was vortexed, and centrifuged at 12,000 x g for 5 min. The aqueous phase containing the RNA was transferred to a separate tube and 0.5 mL of isopropanol was added to precipitate the RNA. Samples were incubated on ice for 10 min and centrifuged at 12,000 x g at 4°C for 10 min. The supernatant was discarded, and the pellet was washed twice with 1 ml 80% ethanol. The RNA pellet was air-dried and re-suspended in 100 μl molecular-grade
C57 and Htr2a EGFP-CreERT2 mice were treated (i.p.) with vehicle or 0.5 mg/kg LSD for 5 days. Twenty-four hr later whole cortex was dissected for Western blot. Receptor purification and subsequent immunoblotting were conducted according to our previously published procedure. The whole cortex was homogenized with a Dounce homogenizer in membrane buffer [50 mM Tris (pH 7.4), 0.1 mM EDTA, 5% glycerol, and with protease and phosphatase inhibitors]. Tissue lysates were centrifuged at 1,000 x g for 10 min at 4°C. The supernatant was collected and centrifuged at 25,000 x g for 15 min at 4°C. The pellets were washed once with membrane buffer followed by re-centrifugation. Pellets (crude membranes) were lysed with lysis buffer [50 mM Tris (pH 7.4), 150 mM NaCl, 10% glycerol, 1% NP-40, 0.5% Nadeoxycholate, 0.5% CHAPS and protease inhibitors] for 1 hr on a rotating mixer at 4°C. Detergent-soluble proteins (supernatants) were collected following centrifugation at 12,500 x g for 20 min at 4°C. The supernatant was incubated at least 2 hr with 30 μl package volume of Wheat-germ (WGA) -conjugated agarose beads (Vector Laboratories, Inc., Burlingame, CA) on a rotating mixer at 4°C and washed 3 times with lysis buffer. WGA-bound proteins were eluted with 50 μl 2x Laemmli buffer [2% SDS, 0.25 M Tris (pH 7.4), 50% glycerol, and 0.01% Bromophenol Blue]. Samples were resolved with 4-12% SDS-PAGE (Life Technologies) and transferred onto polyvinylidene fluoride membranes. Membranes were incubated overnight with antirabbit-HTR2A antibody (1:500, Neuromics, #RA24288) at 4°C followed by horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (Jackson Immunoresearch) and finally reacted with enhanced chemiluminescence reagents. Images were captured with the ChemiDoc Image system (Bio-Rad Laboratories, Hercules, CA). Membranes were stripped and re-probed for β-actin (1:1000, Sigma, #A3854). Intensities of HTR2A were normalized against β-actin in the same lane. Samples from vehicle-treated Htr2a +/+ mice served as the control and were designated as 100%; other samples were normalized against it. The specificity of this HTR2A antibody has been confirmed in HTR2A KO mice in our previous study.
These responses were filmed over 30 min during the open field studies following administration of drugs. Behavioral scoring for HTRs was conducted by researchers blinded to the sex, genotype, and treatment conditions of the mice. Scoring of grooming and retrograde walking were performed using the TopScan program (version 3; CleverSys Inc., Reston, VA). The results are expressed as the numbers of HTRs, duration of grooming (sec), and numbers of retrograde walking events.
PPI of the acoustic startle response was conducted in SR-LAB apparati (San Diego Instruments, San Diego, CA). Mice received (i.p.) the vehicle, 0.3 mg/kg LSD, 1 mg/kg DOI, or 1 mg/kg psilocin and were placed into the PPI chambers. Following habituation to a white-noise background (64 dB), testing began 10 min later. Each test was composed of 42 trials (6 null trials + 18 pulse-alone trials+ 18 prepulsepulse trials). Null trials consisted of the white noise background. Pulse trials were composed of 40 msec bursts of 120 dB white noise (see below). Prepulse-pulse trials were comprised of 6 trials with the 20 msec prepulse stimuli (4, 8, and 12 dB above the white-noise background) that were followed 100 msec later with the pulse stimulus (120 dB). Testing commenced with 10 pulse-alone trials followed by combinations of the prepulse-pulse and null trials, and it ended with 10 pulse-alone trials#'MM;'*<.'D<4D?4<>,E' were performed in slices perfused with aCSF at 32°C (flow rate = 2.5 ml/min). Recordings were made using a Multiclamp 700B amplifier (Molecular Devices). Signals were digitized at 8 kHz using a Digidata 1440A (Molecular Devices), and filtered at 4 kHz. To measure the intrinsic membrane properties of mPFC HTR2A + neurons, whole-cell recordings were conducted in current-clamp mode and series resistance (10-30 MΩ) was monitored throughout the recording session with neurons discarded if resistance changed by >20%. DCZ (250 nM) was bath-applied after 2.5 min of baseline recording and monitored for responses. Data were acquired and analyzed using pCLAMP11 (Molecular Devices). Firing rates and resting membrane potentials were analyzed during baseline and after bath application of DCZ.
All behavioral data analyses were performed with the IBM SPSS Statistics 27 programs (IBM, Chicago, IL). The results were presented as means and standard errors of the mean (SEMs). Student's t-test, one-or two-way ANOVA, repeated measures ANOVA (RMANOVA), or analyses of covariance (ANCOVA) were used, and these analyses were followed by Bonferroni corrected pair-wise acomparisons. When Levene's test of homogeneity of variabce was violated, the data were analysed by Mann-Whitney U or Wilcoxin tests for two groups or by Kruskal-Wallis tests for more than two groups, the latter was followed with Bonferroni tests. A p<0.05 was taken as significant. The results were graphed using GraphPad Prism (San Diego, CA). The results of the binding assay and western blot were analyzed by GraphPad Prism with t-test and two-way ANOVA, respectively. The electrophysiology results were analyzed by GraphPad Prism with paired t-tests and one-sample t-tests. The symbol "*" is for genotype difference (black); "+" is for within C57 effect (blue); "^" for within Htr2a EGFP-CreET2 effect (green), and "#" for overall treatment effects (purple). All data are represented as means ± SEMs. (B) Real-time qPCR was used to determine Htr2a mRNA levels from the whole cortex of C57 mice (Htr2a +/+ ), Htr2a EGFP-CreERT2/+ , and Htr2a EGFP-CreERT2/EGFP-CreERT2 mice. Data are means ± SEMs (n=4 samples/genotype). One-way ANOVA [(F(2,9)=129.4, p<0.0001] followed by Tukey's post-hoc test showed ***p<0.001, C57 vs. Htr2a EGFP-CreERT2/+ ; ****p<0.0001, C57 vs. Htr2a EGFP-CreERT2/EGFP-CreERT2 . (C) LSD-mediated HTR2A downregulation: Mice (C57 and Htr2a EGFP-CreERT2 ) were injected with vehicle or LSD (0.5 mg/kg, i.p.) for 5 consecutive days and then were euthanized 24 hr later. The whole cortices were dissected followed by receptor purification using to the serine residue of the human HTR2A and this was followed at the murine 452 residue by an EGFP insertion; the IRES-Cre was inserted after the exon 3 coding sequence. In the diagrams, the blue boxes represent exons of the murine Htr2a gene; the green box is Egfp; the red box is the stop codon. The gray box is the IRES (internal ribosome entry site) followed by a pink box for Cre or CreERT2 recombinase. The white box is the A242S point mutation. The schematic was created with. (B) HTR2A distribution in the brain among four mouse lines (Htr2a +/+ , Htr2a Cre/+ , Htr2a EGFP-CreERT2/+ , and Htr2a A242S-EGFP-Cre/+ ). For this study, mice were perfused, and the brain sections were immunostained with an anti-HTR2A antibody to visualize the HTR2A distribution. (C) The HTR2A receptor from four mouse lines (Htr2a +/+ , Htr2a Cre/Cre , Htr2a EGFP- CreERT2/EGFP-CreERT2 , and Htr2a A242S-EGFP-Cre/A242S-EGFP-Cre ). Here, mice were euthanized and then cortices were dissected followed by receptor purification using Wheat-Germ beads pull-down. Western blot against anti-HTR2A antibody to detect receptor expression level. In Htr2a +/+ and Htr2a Cre/Cre mice, the HTR2A was detected between 50-75 KD; for EGFP insertion lines, the HTR2A-EGFP-CT fusion protein was detected around 100 KD. Example electrophysiological trace of a HRT2A + neuron expressing DIO-HA-hM3Dq-IRES-mCitrine during bath application of DCZ. Htr2a EGFP-CreERT2/+ x Ai9 or C57 x Ai9 (Htr2a +/+ x Ai9) mice were treated with vehicle (corn oil) or tamoxifen (100 mg/kg) at p39 to p42 with 4 injections and then brains were collected at p56. A whole brain mapping was conducted to determine whether CreERT2 recombinase turned on tdTomato protein expression after TAM treatment. The brain sections were stained with DAPI and then images were rendered by an Olympus VS120 slide scanner. Experiments were conducted with 4 animals with similar results. In contrast to the frontal pole, the general distribution of HTR2A-EGFP-CT in the medial prefrontal cortex--which contains a defined infralimbic cortex (IL, areas 25)--was similar to that seen in the frontal pole with one exception. Here, the L5a band was not present in the IL or in the ventral prelimbic cortex (PL, area 32) (Suppl. #1 Figure). Ventral to the IL, the medial orbital cortex also contained HTR2A-EGFP-CT. There are no consistent cytoarchitectonic features that distinguish dorsal and ventral prelimbic cortex. However, the dorsal PL was more densely invested than its ventral counterpart with a greater density of acetylcholinesterase (AChE) -positive fibers. We therefore examined adjacent tissue sections stained to reveal AChE and HTR2A-EGFP-CT. The ventral (AChE-weak) zone of the PL did not express HTR2A-EGFP-CT but the dorsal PL did (Suppl.#1 Figure).
As in the frontal cortices, HTR2A-EGFP-CT in areas posterior to the genu of the callosum formed an arching band in L5a. HTR2A-EGFP-CT extended from the cingulate cortex medially to the rhinal sulcus laterally. The HTR2A-EGFP-CT in the postgenual cortex were most dense in L5a but were present also at lower density in L6, with fewer still in L2/3. HTR2A-EGFP-CT labeling of apical dendritic arborizations was seen in L1. Ventral to the rhinal sulcus there was a modest density of HTR2A-EGFP-CT that marked the piriform cortex. Caudal to the piriform cortex, the density of HTR2A-EGFP-CT increased in the perirhinal cortex and was very dense in the lateral, but not the medial entorhinal cortex. At this level the intensity of HTR2A-EGFP-CT staining was low-tomoderate in the primary auditory and visual cortices and it was present also in secondary visual and auditory cortical areas. Anterior olfactory areas. There was a moderate plexus of HTR2A-EGFP-CT-positive elements in the anterior olfactory nucleus, particularly in the ventral aspects. The anteriormost piriform cortex was labeled. Interestingly, HTR2A-EGFP-CT was present in low-tomoderate densities in the lateral olfactory tract embedded in the ventrolateral aspect of the olfactory bulb, but it was very dense in the lateral olfactory tract axons that course to and enter the olfactory bulb. Claustrum and endopiriform cortex. The claustrum exhibited moderate HTR2A-EGFP-CT signal through most of its anteroposterior extent, although labeling was more pronounced in the central core of the claustrum. Ventral to the claustrum, HTR2A-EGFP-CT was present in the ventrally contiguous endopiriform nucleus and in the cortex dorsally adjacent to the claustrum.
Both the dorsal striatum (caudatoputamen) and the ventral striatum (nucleus accumbens and olfactory tubercle) exhibited areas of dense HTR2A-EGFP-CT signal. In the dorsal striatum, small clusters of intense HTR2A-EGFP-CT staining were embedded in a much less dense background area. This pattern was similar to the striatal patch-matrix system, in which small islands of weak acetylcholinesterase (AChE) termed striosomes were intermixed with a surrounding matrix compartment enriched in AChE. We therefore examined sections stained with a µ-opioid receptor (MOR) antibody that selectively stains striosomes. MOR and HTR2A-EGFP-CT were co-localized in very densely filled striosome-like compartments (Figure). In addition, the subcallosal stria, a thin (~50 µm) band of striatal tissue that is an elongated striosome running along the ventral surface of the callosum, was both MOR-and HTR2A-EGFP-CT-positive (Figure). The density of HTR2A-EGFP-CT expression in the matrix (MOR-negative) compartment was observed to differ across the rostrocaudal extent of the dorsal striatum. In the rostral striatum, HTR2A-EGFP-CT-enriched striosomes were present over a diffuse light background of HTR2A-EGFP-CT matrix staining, while more posteriorly in the midstriatum densely-stained HTR2A-EGFP-CT-positive striosomes were seen in a matrix almost devoid of HTR2A-EGFP-CT staining. However, still more caudal, in the tail of the striatum, it was difficult to differentiate GFP in striosomal patches from the background HTR2A-EGFP-CT matrix staining. In the rostral half of the nucleus accumbens, HTR2A-EGFP-CT densely distributed the ventrolateral shell and core regions but was much less dense in the medial shell and septal pole areas. More posteriorly, the medial shell was almost devoid of HTR2A-EGFP-CT, while the lateral shell and core were very densely filled with fibers. These distinctions in intensity of HTR2A-EGFP-CT labeling did not correspond to labeling in different accumbal compartments--such as the core and shell--but rather adopted a simply a lateral-to-medial gradient across the accumbens. The intensity of HTR2A-EGFP-CT in the olfactory tubercle (TUO) paralleled that seen in the accumbens, with the marker of HTR2A-EGFP-CT expression being less dense in the medial than lateral TUO. The lateral TUO was among the most densely-stained areas in the mouse brain.
There was no significant HTR2A-EGFP-CT of either the globus pallidum or the ventral pallidum. Septum, amygdala, and hippocampus. The septum and diagonal band complex lacked significant HTR2A-EGFP-CT. This led to a forebrain staining pattern in which dense GFP staining in the dorsal and ventral striatum flanked the septum and diagonal band, which lacked significant GFP staining. The amygdala displayed a heterogenous pattern of HTR2A-EGFP-CT, with most of the amygdala exhibiting very low or no HTR2A-EGFP-CT while the extreme lateral and medial nuclei were moderately-to-densely distribution. HTR2A-EGFP-CT was dense in the lateral nucleus, but the ventromedially-adjacent basolateral nucleus was devoid of HTR2A-EGFP-CT. There was almost no HTR2A-EGFP-CT in the central nuclei, with weak staining of the posterior basomedial and cortical nuclei. The medial amygdalar nucleus was densely filled with HTR2A-EGFP-CT. In the dorsal hippocampus there was trace expression of HTR2A-EGFP-CT in the pyramidal cell areas and the dentate gyrus. In the ventral (temporal) hippocampus there was a very different pattern, with almost no HTR2A-EGFP-CT in the CA1 and CA2 fields but very intense staining of the ventral aspects of the CA3 field with GFP-labeled pyramidal cell dendrites visible in the stratum oriens. The dentate gyrus did not express significant HTR2A-EGFP-CT.
The HTR2A-EGFP-CT was widely distributed in the forebrain, with particular enrichment in cortical areas and the striatal complex. In contrast, HTR2A-EGFP-CT expression appeared to be progressively less as one moved from the forebrain to diencephalic, midbrain and pontomedullary areas. Thalamus and epithalamus. Most of the thalamus lacked significant HTR2A-EGFP-CT. There were two exceptions to the paucity of the HTR2A-EGFP-CT in the thalamus. The subparafascicular nucleus was filled with a moderate density of GFP fibers, while a light plexus of GFP staining filled the anterior pretectal nuclei. In the epithalamus, the lateral habenula was invested with a light-to-moderate density of HTR2A-EGFP-CT-positive fibers, while in the medial habenula significant HTR2A-EGFP-CT expression was restricted to the dorsomedial tip of the structure. There was no appreciable HTR2A-EGFP-CT in the paraventricular nucleus. Hypothalamus. HTR2A-EGFP-CT was found in only a few hypothalamic areas. The ventromedial nucleus displayed light-to-moderate GFP staining, with trace levels in the dorsomedial nucleus. The zona incerta was lightly filled with HTR2A-EGFP-CT. In the posterior hypothalamus, HTR2A-EGFP-CT was seen in the mammillary bodies. Little HTR2A-EGFP-CT was present in the lateral mammillary nucleus and the ventral premammillary nucleus, but very dense staining was present in the medial and lateral divisions of the mammillary area, as well as, in the dorsal but not ventral tuberomammillary nuclei.
Trace or low levels of diffuse HTR2A-EGFP-CT signal were seen in a numbers of midbrain areas, including the mesencephalic periaqueductal gray, the red nucleus, and the nucleus of Darkschewitz. The density of the HTR2A-EGFP-CT, as reflected by GFP, was low in the interpeduncular nucleus (IPN). There was slightly greater expression in the lateral and dorsal IPN subnuclei than the central and intermediate IPN nuclei (as defined by. Dorsally in the midbrain, low HTR2A-EGFP-CT labeling was present in the superior colliculus, with the greatest density in the intermediate gray layer. In the intermediate gray staining inhomogeneities contributed to a patchy appearance. GFP staining density was less intense in the inferior colliculus. Weak GFP expression was observed starting in the mesencephalic periaqueductal gray and extending caudally into the rostral pontine periaqueductal gray.
Pons. Across the basilar pons, areas of dense or very dense GFP staining were interdigitated with weak or absent HTR2A-EGFP-CT. In the dorsomedial and dorsolateral nucleistaining was dense, while the ventral pontine cells displayed much weaker staining. In addition, there was moderate staining in the tegmental reticular nucleus (of Bechterew) dorsally. HTR2A-EGFP-CT was moderately dense in the dorsal tegmental nucleus; the surrounding central gray showing significantly less intense staining. In the parabrachial complex, the dorsal parabrachial and Kolliker-Fuse nuclei were weakly positive for HTR2A-EGFP-CT, with trace amounts of the fluorophore observed in the ventral parabrachial nuclei. The paralemniscal nucleus showed light-to-moderate GFP, and the nucleus ambiguus was lightly stained. Prominent by virtue of absence was HTR2A-EGFP-CT expression in the raphe nuclei. Thus, there was no consistent staining of the dorsal or median raphe nuclei or of the raphe obscurus, while trace HTR2A-EGFP-CT was seen around the lateral aspects of the raphe magnus. Cranial nerve nuclei exhibited a moderate density HTR2A-EGFP-CT, prominently including the fifth (V, trigeminal) and seventh (VII, facial) cranial nerve nuclei.
A few medullary areas expressed HTR2A-EGFP-CT. Most prominent among these was the inferior olivary complex (IO). Although staining throughout the IO was moderate-to-dense, there were subnuclear variations in GFP density, including a (relatively) high density of staining in subnucleus A of the medial nucleus, moderate staining in the principal nucleus, and less intense staining in the cap of Kooy of the medial nucleus [following the nomenclature of].
No specific HTR2A-EGFP-CT was observed in either the cerebellar cortex or the deep nuclei. White matter. Despite the dense expression of HTR2A-EGFP-CT in L5 cortical neurons, which gave rise to extensive descending myelinated projections as well as some interhemispheric connections, there was no significant GFP labeling of the corpus callosum. Similarly, neither the anterior commissure, nor the internal capsule were labeled. However, in the forebrain the lateral olfactory tract was lightly labeled. By contrast, the lateral olfactory tract outside of the corpus of the brain was very intensely labeled. More posteriorly in the brain several myelinated fiber areas expressed some HTR2A-EGFP-CT, although in no case was such staining dense. In the posterior thalamus, the fasciculus retroflexus exhibited light-to-moderate HTR2A-EGFP-CT and could be traced from the habenula to the interpeduncular nucleus. In addition, low levels of HTR2A-EGFP-CT were observed in the medial accessory oculomotor tract, and a small number of moderately-stained HTR2A-EGFP-CT-positive axonal clusters were visible in the lateral lemniscus. Finally, moderate HTR2A-EGFP-CT was present in the axons of the facial nerves.
Create a free account to open full-text PDFs.
Preller, K. H., Burt, J. B., Adkinson, B. et al. · eLife (2018)
Shao, L-X,, Liao, C., Gregg, I. et al. · Neuron (2021)
Vargas, M. V., Dunlap, L. E., Dong, C. et al. · Science (2023)
Vollenweider, F. X., Vollenweider-Scherpenhuyzen, M. F. I., Bäbler, A. et al. · NeuroReport (1998)
Doss, M. K., Madden, M. B., Gaddis, A. et al. · Brain (2021)
Papers in Blossom that reference this study
Yacoub, J., Bray, E., Bayyat, J. et al. · Biorxiv (2026)
Fleury, S., Nautiyal, K. M. · Biorxiv (2024)
Ekins, T. G., Brooks, I., Kailasa, S. et al. · Biorxiv (2023)